GitOps – the future of DevOps?

Get valuable insights on how GitOps aids DevOps and investigate the phenomenon from multiple angles in order to achieve a deeper understanding of delivering products and services at high velocity.

One of the key drivers of the software development process is the ability to eliminate the separation between development (Devs) and operations teams (Ops). For that reason, companies adopt a DevOps culture. It allows them to deliver products and services at high velocity and with fewer issues to tackle. The numbers show that the global DevOps market is expected to reach $58 billion by 2030 while it was valued only at $6.7 billion in 2020.

DevOps has been the methodology of choice for companies willing to implement and improve their products fast for over a decade now. Yet, while DevOps offers an environment of shared transparency, responsibility, and productivity, it has its downsides. Out of the need to compensate for the challenges presented by DevOps, GitOps was born. Further, one should explore GitOps from various angles so as to prove or disprove the following hypothesis – GitOps is better than DevOps.

GitOps in a nutshell

GitOps is currently in the stage of its germination. The numbers suggest the phenomenon will reach $3.85 billion by 2023 from $1.85 billion in 2018. It has an expected Compound Annual Growth Rate (CAGR) of 18.5% in five years (see Fig. 1).Figure 1. Continuous delivery market dynamics by Development Mode

This shows that for GitOps, everything is still up ahead. Thus far, it all depends upon the ability of the technology to prove its worth and to bring as many advantages as possible. With the background established, the key question remains: what is GitOps?

GitOps model

When it comes to the GitOps model, there are three key elements to consider:

1. Infrastructure as Code (IaC). As it was mentioned before, GitOps uses Git, which is a Git repository. In this case, the repository is presented as the source of truth for all the definitions within the system’s infrastructure. All the changes employed are tracked and stored within the Git repository. IaC is the instrument for keeping all the configurations as code.
2. Merge Requests (MRs). The GitOps approach is about using MRs as change tools for infrastructure updates. It serves as an audit log and allows teams to collaborate around formal approvals for reviews and comments.
3. Continuous integration and Continuous Delivery (CI/CD). The GitOps model includes CI/CD so as to enable a changed environment every time new code is merged. Essentially, CI/CD helps manage the GitOps automation.

These elements represent the GitOps model and illustrate how it operates with DevOps and Git techniques.

GitOps principles

With methodologies like DevOps and GitOps, foundational principles are always constituting a broader culture. For GitOps, the following principles should be mentioned:

• Version control. With a Git repository being the source of truth definitions, it is easier to make changes if required.
• Declarative infrastructure. This allows developers to emphasize application development without the necessity to deal with the manual handling of runtimes and logistics of deployment.
• Software agents to ensure accuracy. These agents help to automatically deal with potential issues if the actual state does not match the desired state.
• Automated change approvals. This principle enables the automatic deployment of any modifications within the system’s infrastructure.

GitOps principles provide valuable feedback to the software development team and make various processes linked to the system’s infrastructure seamless and automated.

GitOps workflow

Along with the GitOps model and GitOps principles, there is a particular GitOps workflow to account for. There are four stages within the workflow:

Stage 1: The Git repository is used for storing the application configuration and code.

Stage 2: The CI/CD pipeline is established to build, test, and deploy the application.

Stage 3: The application resources within the target environment are managed via a deployment tool.

Stage 4: The team tracks the application’s performance through the monitoring system.

Following these phases constitutes a GitOps workflow. Piecing all the elements together, one can see GitOps as a new methodology that uses the best of DevOps practices and adopts them into an environment allowing better automation, monitoring, and feedback.

Why is GitOps significant?

At this moment, it is very clear that GitOps offers advantages of boosted automation and more effective control measures. However, the list does not stop with those two factors. Respectively, offering a more comprehensive view of benefits brought by GitOps, one cannot forget the following:

• Enhanced security. Using a single platform for change management minimizes downtime, allowing developers to work without disruptions and compromise.
• Improved collaboration and productivity. GitOps optimizes the CI/CD pipeline, simplifying various processes and improving oversight, while allowing every team player to understand and manage individual processes.
• Less manual input. With GipOps, teams do not need to spend time on manual tasks and can focus on software development, all due to automation brought by the technology.
• Rapid development. Practices of GitOps enable teams to use open-source instruments to perform integration and deployment, which results in faster releases with almost no delays.
• Boosted reliability. The GitOps model and workflow allow teams to better identify errors through the possibility of rolling back the new version to the previous one.
• Being vendor-neutral. Most practices and instruments integrated into GitOps are open-source and do not need any third parties or vendors.

• Network narrowing. Service providers can use GitOps to create systems allowing users to pay only for the traffic they use or need. As a result, it can mean better deals like premium pricing for video streaming and/or lower pricing for Internet of Things (IoT) services.
• Smart city services. Managing smart city services is a complicated task due to several benefits that need to be managed simultaneously. With such an opportunity comes a chance to reduce network load and help manage operational complexities.
• Network congestion. As areas become more urbanized, service providers are seeking solutions for network congestions. This issue was apparent with the adoption of 4G. Often, when thousands of people use the network simultaneously, the overall bandwidth can drop. With the help of GitOps and cloud engineering, the issue can be resolved by making networks more sustainable.

There are not only theoretical benefits of GitOps. The use cases mentioned above show how the methodology can aid real projects. Undoubtedly, the more GitOps is developed, the more practical applications will emerge.

GitOps potential issues to consider

The DevOps market is expected to experience a CAGR of a staggering 20 percent from 2022 to 2030. Essentially, it means that with further integration of cloud technologies and the adoption of fast application delivery methods, like GitOps, the market will grow and bring billions of dollars (see Fig. 2).Figure 2. U.S. development to operations market size

However, like with every technology, old or new, there are particular elements to consider. Understanding the potential GitOps issues further makes a difference between properly applied GitOps and it doing more harm than benefit:

• CI/CD separation. In contrast to DevOps, GitOps does not make deployment of CI/CD a last stage in the pipeline. The process is triggered when a Git push is enabled, which leads to the separation of CI and CD. Yet, practically, such separation is hard to achieve.
• Bounded observability. GitOps uses the basic abstraction level by creating a condition of limited visibility for product managers. Often, only developers have the complete information on Git hashes and commits.
• Challenges with rollbacks. Multiple past commits emerge with different software development teams operating the same GitOps framework. This creates issues with setting the rollbacks. Usually, to avoid this problem, only one team should use GitOps.
• Concerns with configurations. Sometimes, a given architecture might require multiple environments. With GitOps, it might be challenging to manage various environments. It happens because everything stored in a single Git source requires a new Git push to be executed every time change is imposed on any of the environments.
• Complex in auditing. If a company develops a large and complex application using GItOps, it can be hard to audit the product. Namely, such an application might have more than one repository, multiple configuration files, and hundreds of data points. There is an option to enable additional observability tooling. However, it contradicts the very nature of GitOps – eliminating the boundary between Devs and Ops.

The issues mentioned above illustrate that GitOps is not a panacea. There are still considerations to be had, and aspects to be accounted for. Still, when done properly, GitOps can bring more simplicity to the deployment process.

Is GitOps better than DevOps?

At this point, there is enough information to ascertain whether GitOps is better than DevOps. But, in terms of GitOps vs. DevOps, you cannot say that the latter is better than the former, and you cannot suggest the opposite either. It might seem paradoxical, but it appears to be true. Instead of determining which approach is best, the more accurate question should be about how GitOps makes DevOps better.

Keeping all the insights mentioned above in mind, one can say that GitOps and DevOps were put together to create a unique approach better than GitOps and DevOps standing alone. Basically, GitOps can boost DevOps’s productivity. To illustrate, after experimenting with various infrastructure configurations, a team of Devs and Ops can appeal to a Git history to revert any given changes in order to restore a good state.

Essentially, this grants teams an undo button that saves time and work. Such an opportunity is possible because GitOps can be used to incorporate Git through the ensured software delivery process, making DevOps easier to orchestrate. After all, both GitOps and DevOps want the same thing: to eliminate the boundary between development and operations while making software delivery faster, smoother, and more reliable. Furthermore, GitOp Kubernetes and Openshift GitOps can be applied to the DevOps framework to automate pipeline workflows and achieve continuous deployment.

The next step in the GitOps and DevOps evolution

GitOps can aid DevOps by bringing more automation and granting the opportunity to reverse any changes easily. Coupled together, these two methodologies can go a long way. However, when it comes to a paradigmatic shift in faster software development and deployment, one cannot avoid mentioning cloud engineering. The U.S. cloud computing market is currently valued at $369 billion and is expected to skyrocket shortly (see Fig. 3).Figure 3. U.S. cloud computing market size

The cloud engineering platform is predicted to be the climax of everything. Devs and Ops learned this while developing and delivering software. The rate of innovation that includes modern cloud technologies is truly immense. This up and coming philosophy dictates that developers must be equipped with elements allowing them to self-serve and be less reliant on external tools.

What is more, cloud engineering helps tackle one of the biggest concerns developers have – how to manage to scale? Any infrastructure change requires input or change through a CI/CD pipeline. With new features added with cloud engineering, developers are given an underlying language-neutral model allowing them to write packages that can be used across different ecosystems and programming languages. Cloud engineering is the next step of evolution in DevOps because it optimizes development across different environments.

Wrapping up

GitOps is neither better nor worse than DevOps. Furthermore, GitOps without DevOps does not have a future. The scope needs to be changed and directed at further investigating how the approach can aid DevOps. It is unlikely that GitOps will become something bigger than DevOps. It is more likely that working together, the methodologies will give rise to the approach of redefining the aspects, like a collaboration between Devs and Ops, as well as bringing automation into the picture. Like with any given technological solution, switching to the GitOps-only approach will not fix all the issues with deployment. However, coupling GitOps with DevOps will bring you closer to a faster, smoother, and sustainable deployment and delivery.

If you want Avenga to help you with DevOps and GitOps in software development, feel free to contact us.